Skip to main content

​
Langchain Service Overview

The MOOD MNKY Langchain service provides a comprehensive set of APIs for building AI-powered applications with language models. This service enables you to create sophisticated AI workflows, manage language model interactions, process documents, and build intelligent applications with minimal development effort.

​
Key Features

  • Language Model Integration - Interact with various language models through a unified API
  • Chain Management - Create, configure, and run complex reasoning chains
  • Memory Systems - Implement conversation history and contextual memory
  • Document Processing - Upload, process, and retrieve documents using vector embeddings
  • Agents - Deploy autonomous agents that can use tools and make decisions
  • Prompt Management - Create, iterate on, and optimize prompts for different use cases

​
Getting Started

To start using the Langchain service, you’ll need:
  1. API Credentials - Obtain your API key from the Developer Portal
  2. Service Endpoint - Connect to the right environment:
    • Development: http://localhost:8000
    • Production: https://langchain.moodmnky.com

​
Base URL

All API endpoints are relative to the base URL: 
https://langchain.moodmnky.com/api

​
Authentication

All requests require an API key passed in the x-api-key header: 
curl "https://langchain.moodmnky.com/api/models" \
  -H "x-api-key: your_api_key"

​
Service Status

Check the current status of the Langchain service: 
curl "https://langchain.moodmnky.com/api/health" \
  -H "x-api-key: your_api_key"
Response: 
{
  "status": "healthy",
  "version": "1.2.3",
  "uptime": "3d 4h 12m",
  "services": {
    "database": "connected",
    "vector_store": "connected",
    "llm_services": "connected"
  }
}

​
Core Concepts

​
Chains

Chains are sequences of operations that combine language models with other components like memory, document retrieval, or tool usage. They enable you to create complex workflows while maintaining a simple interface. Learn more about Chains →

​
Memory

Memory systems allow your applications to maintain conversation history, remember user preferences, and provide contextual awareness across interactions. Learn more about Memory →

​
Documents

The document processing system lets you upload, process, and retrieve documents using vector embeddings, enabling knowledge-based applications and retrieval-augmented generation. Learn more about Documents →

​
Common Workflows

​
Build a Conversational AI

// Create a conversation chain
const response = await axios.post(
  'https://langchain.moodmnky.com/api/chains',
  {
    name: 'Customer Support Assistant',
    type: 'conversation_chain',
    llm: {
      provider: 'ollama',
      model_name: 'llama2',
      temperature: 0.7
    },
    memory: {
      type: 'conversation_buffer',
      memory_key: 'chat_history',
      return_messages: true
    },
    prompt_template: `You are a helpful customer support agent for MOOD MNKY.
    
Current conversation:
{chat_history}

User: {input}
AI: `
  },
  {
    headers: {
      'x-api-key': 'your_api_key',
      'Content-Type': 'application/json'
    }
  }
);

// Get the chain ID from the response
const chainId = response.data.id;

// Run the conversation chain
const chatResponse = await axios.post(
  `https://langchain.moodmnky.com/api/chains/${chainId}/run`,
  {
    input: 'How do I create a custom fragrance?'
  },
  {
    headers: {
      'x-api-key': 'your_api_key',
      'Content-Type': 'application/json'
    }
  }
);

console.log(chatResponse.data.output);

​
Build a Document Q&A System

// Create a retrieval QA chain
const response = await axios.post(
  'https://langchain.moodmnky.com/api/chains',
  {
    name: 'Product Documentation QA',
    type: 'retrieval_qa',
    llm: {
      provider: 'ollama',
      model_name: 'llama2',
      temperature: 0.2
    },
    retriever: {
      vector_store_id: 'vector-store-1234',
      search_type: 'similarity',
      k: 5
    },
    prompt_template: `You are an AI assistant for MOOD MNKY product documentation.
    
Answer the question based only on the following context:
{context}

Question: {question}
Answer: `
  },
  {
    headers: {
      'x-api-key': 'your_api_key',
      'Content-Type': 'application/json'
    }
  }
);

// Get the chain ID from the response
const chainId = response.data.id;

// Run the QA chain
const qaResponse = await axios.post(
  `https://langchain.moodmnky.com/api/chains/${chainId}/run`,
  {
    question: 'What ingredients are in the Calm Fragrance Blend?'
  },
  {
    headers: {
      'x-api-key': 'your_api_key',
      'Content-Type': 'application/json'
    }
  }
);

console.log(qaResponse.data.output);

​
Integration with Other Services

The Langchain service integrates seamlessly with other MOOD MNKY services:
  • Ollama Service - Used for language model inference
  • Flowise Service - For visual workflow creation and deployment
  • n8n Service - For automation and integration with external systems

​
Best Practices

  1. Use the Right Chain Type
    • Choose the appropriate chain type for your use case
    • Use conversation chains for chat applications
    • Use retrieval QA chains for knowledge-based applications
    • Use sequential chains for multi-step workflows
  2. Optimize Prompts
    • Create clear, specific prompt templates
    • Include appropriate context and instructions
    • Test and iterate on prompts for best results
  3. Manage Memory Effectively
    • Choose the right memory type for your application
    • Limit conversation history length to manage tokens
    • Consider using summarization memory for long conversations
  4. Document Processing
    • Use appropriate chunk sizes for your content
    • Include comprehensive metadata for better retrieval
    • Consider document structure when designing your system

​
Rate Limits

PlanRequests per MinuteRequests per DayTokens per Request
Development6010,00016,000
Basic12050,00032,000
Professional300250,00064,000
EnterpriseCustomCustomCustom
Exceed these limits and your request will receive a 429 Too Many Requests response.

​
Support & Resources